Abstract
Let be the Hodges-Lehmann estimator (based on an appropriate one-sample rank statistic) of the center θ of a symmetric and absolutely continuous distribution. Let measure-ments be rounded off, to, say the nearest integers, and
the corresponding estimator based on the grouped data (ties being handled by the average scores method). The following two results hold:
(i) With probability one.
(ii) for a general θ , and
if θ is an integer or zero
For any given α , a confidence interval for θ , with confidence coefficient at least 1-α , is constructed. Applications to Chemistry, Weight Training Data and Head Circumference Data illustrate.